Head of Robotics Data Acquisition & Strategy

About the team and the role

The Data Strategy function decides what data Mind buys, from whom, and why. Robot foundation models are only as good as the data they learn from, and data is one of the largest investments we make in our research program. This function turns that spend into a deliberate, measured portfolio: the right modalities, tasks, embodiments, and environments, sourced from the right partners at the right price, and landing training-ready for our research teams.

As Head of Robotics Data Acquisition & Strategy, you will own Mind's data purchase plan, vendor selection, and vendor management end to end, and you will report directly to Daniel Burrows, Head of Data. Done well, we will know the impact of each additional hour of data we buy on model performance, and allocate our budget accordingly. You will also maintain deep market awareness of the robotics data landscape, monitoring collection methodologies, industry pricing, and emerging capabilities across the sector to proactively recommend updates to our data strategy. You'll travel to vendors and conferences and bring what you learn back into our research org. This is a rare chance to define a new function at the center of a physical AI company, with direct influence over the research roadmap and the bets the company makes.

Responsibilities

• Own the data purchase plan. Build and maintain Mind's annual and quarterly data acquisition plan: hours by modality (teleoperation, egocentric human video, handheld/wearable capture, motion capture, simulation and synthetic data), task, embodiment, and environment, tied to the research roadmap, model evaluation targets, and budget.

• Measure the marginal value of data. Partner with research to design and run data ablations, scaling studies, and data-mixture experiments that quantify how each additional hour of a given data type moves policy performance on held-out evaluations. Translate results into cost per unit of model improvement and use them to reallocate spend.

• Run vendor selection end to end. Run market scans, RFPs, and paid pilot collections with defined acceptance criteria. Conduct technical diligence on vendors' collection rigs, sensor configurations, calibration and synchronization, metadata, annotation pipelines, and their ability to scale.

• Manage the vendor portfolio. Negotiate contracts and SLAs; set and enforce quality specifications (sensor sync, frame rates, calibration, annotation accuracy, task coverage); track throughput, cost per usable hour, and delivery cadence; and expand, renegotiate, or exit partners based on measured value.

• Define data specifications with research and data infrastructure. Set the formats, schemas, sensor modalities, labeling taxonomies, and acceptance tests that purchased data must meet so it lands training-ready, and close the loop on quality issues found downstream.

• Make build-versus-buy calls. Weigh external purchases against Mind's own in-house and factory-floor collection, and recommend where each dollar of data budget creates the most value.

• Be our eyes and ears in the market. Maintain a living map of the robotics data landscape: who is collecting what, by which methods, at what price, quality, and capacity, and which new collection approaches are emerging.

• Bring the outside in. Attend major robotics and ML conferences (e.g. CoRL, RSS, ICRA, CVPR, NeurIPS), visit vendor and field sites, and synthesize what you learn into regular briefings and recommendations that shape the research org's priorities.

• Manage data rights and risk. Partner with Legal and Finance on data licensing, IP and usage rights, privacy and consent for human-subject data, and contract terms, so every dataset we buy is one we can use.

• Set direction for the Data Acquisition function. Advise executive staff on data investment bets, establish the function's operating cadence and metrics, and hire as scope grows.

Requirements

• Bachelor's or Master's degree in Computer Science, Robotics, Electrical or Mechanical Engineering, Statistics, or a closely related quantitative field, or equivalent experience.

• Deep familiarity with training data for machine learning in robotics, embodied AI, autonomous vehicles, or a comparable physical-world domain, including how data scale, diversity, and quality affect model performance.

• Demonstrated experience designing quantitative experiments (ablations, scaling studies, or A/B analyses) and using the results to drive investment decisions.

• Hands-on fluency in Python and SQL sufficient to analyze datasets, evaluation results, and vendor metrics yourself.

• Working knowledge of physical data-collection methods and sensor stacks: teleoperation, egocentric and multi-camera video, motion capture, IMU and force/torque sensing, calibration, and time synchronization.

• Track record of owning a multi-million-dollar data, vendor, or procurement budget and managing five or more external vendors concurrently, from sourcing and RFP through contract, SLA, and performance management.

• An existing network across the robotics and AI data ecosystem (data vendors, research labs, and industry peers).

• Ability to set direction in an ambiguous, fast-moving domain and to communicate clearly with both researchers and executives, in writing and in person.

• Familiarity with data licensing, IP, and privacy/consent requirements for human-subject data.

• Willingness to travel approximately 20% of the time, including international travel to vendor sites and conferences.